Google announced the rollout of a music-generation capability inside the Gemini app powered by DeepMind’s Lyria 3 model, currently available in beta. The feature enables users to generate original songs from natural language prompts, producing a 30-second track with lyrics and cover art, and can also create music based on uploaded photos or videos.
Lyria 3 delivers more realistic compositions and allows users to control elements such as style, vocals, and tempo. Google is also expanding access to its Dream Track tool for YouTube creators worldwide, extending availability beyond the United States.
All generated audio will include SynthID watermarking to identify AI-produced content, and Gemini will support detection of AI-generated music. The feature is launching globally for adult users across multiple major languages.
